]> err.no Git - linux-2.6/commit
sched: fair-group: SMP-nice for group scheduling
authorPeter Zijlstra <a.p.zijlstra@chello.nl>
Sat, 19 Apr 2008 17:45:00 +0000 (19:45 +0200)
committerIngo Molnar <mingo@elte.hu>
Sat, 19 Apr 2008 17:45:00 +0000 (19:45 +0200)
commit18d95a2832c1392a2d63227a7a6d433cb9f2037e
treefa85b700aa3caac5b1309edd8e31d9b957957a83
parent1d3504fcf5606579d60b649d19f44b3871c1ddae
sched: fair-group: SMP-nice for group scheduling

Implement SMP nice support for the full group hierarchy.

On each load-balance action, compile a sched_domain wide view of the full
task_group tree. We compute the domain wide view when walking down the
hierarchy, and readjust the weights when walking back up.

After collecting and readjusting the domain wide view, we try to balance the
tasks within the task_groups. The current approach is a naively balance each
task group until we've moved the targeted amount of load.

Inspired by Srivatsa Vaddsgiri's previous code and Abhishek Chandra's H-SMP
paper.

XXX: there will be some numerical issues due to the limited nature of
     SCHED_LOAD_SCALE wrt to representing a task_groups influence on the
     total weight. When the tree is deep enough, or the task weight small
     enough, we'll run out of bits.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
CC: Abhishek Chandra <chandra@cs.umn.edu>
CC: Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
include/linux/sched.h
kernel/sched.c
kernel/sched_fair.c
kernel/sched_rt.c